Spreading of contents that infringe copyright via networks is attracting more and more attention. Correct identification of the contents that infringe copyright is the basis for managing (for example, filtering or screening) the contents spread via networks.
A current secure content identification mechanism (SCIDM) is a content-based identification mechanism based on fingerprint extraction (referred to as a fingerprint-based identification mechanism in the following). An SCIDM-based content identification system manages an SCIDM server (also referred to as a Content Identification Manager (CIM)). A protected content is registered with the CIM, and the CIM extracts a fingerprint from the protected content and stores the fingerprint, and meanwhile, the CIM stores related attribute information (for example, copyright ownership information or a copyright protection rules) of the protected content. In addition, the SCIDM-based content identification system also defines an SCIDM client (also referred to as a Monitor Entity (ME)). The SCIDM client is adapted to monitor whether a content passing through or sent to an entity such as a monitor gateway, a user terminal or a content sharing website infringes copyright. During the monitoring process, the SCIDM client extracts a fingerprint from the received content, and sends the extracted fingerprint to the SCIDM server; and the SCIDM server locally searches for a fingerprint of a protected content matching the extracted fingerprint. If the SCIDM server finds the fingerprint of the protected content matching the extracted fingerprint, the SCIDM server returns related content attributes to the SCIDM client, and the SCIDM client manages (for example, filters or screens) corresponding contents according to the content attributes.
In the implementation of the present invention, the inventor found that in the existing fingerprint-based identification mechanism, as the SCIDM client extracts the fingerprint of the content and the SCIDM server performs content identification operations (such as searching and matching operations) according to the fingerprint of the content, a lot of computing resources are consumed. As a great number of audio and video contents exist in the network, a heavy load is imposed on the SCIDM client and the SCIDM server, resulting in a reduction in the identification efficiency.